17th World Conference on Nondestructive Testing, 25-28 Oct 2008, Shanghai, China
An Automatic Guided Wave Pulse Position Modulation System Using Steel Pipes as a Communication Channel for Flood Detection in Steel Offshore Oilrigs Rito MIJAREZ 1, Fernando MARTINEZ 1, Patrick GAYDECKI 2 1
Gerencia de Control e Instrumentación, Instituto de Investigaciones Eléctricas, 62490, Cuernavaca, Morelos, Mexico; e-mail:
[email protected] 2 School of Electrical and Electronic Engineering, The University of Manchester, PO Box 88, Manchester M60 1QD, UK; e-mail:
[email protected]
Abstract An automatic guided wave pulse position modulation system, using steel tubes as communication channel, for detecting flooding in the hollow sub-sea structures of offshore oilrigs is presented. The system employs a smart sensor and modulator that transmit guided wave pulses in the range of 40 kHz and is composed of a PZT element, electronics and batteries. The demodulator is made of a PZT ultrasound transducer, a DSP board and a module based on a microcontroller, which performs automatic detection of guided wave energy packets. Experiments carried out in steel tubes 1.5 m x 0.27 m x 2 mm, in air, have successfully detected automatically guided wave encoded information. Keywords: Offshore platforms monitoring, Guided waves, Pulse Position Modulation 1. Introduction The health of underwater supporting steel structures of fixed steel offshore platforms, normally made from sealed steel tubes filled with air, is an important issue as failure is extremely costly in terms of money and possibly human life. The infiltration of seawater via throughthickness cracks into the sealed crossbeams is a crucial matter, and is used as a basis for the development of an inspection method known as flooded member detection (FMD). This method provides great appeal as a simple, reliable and efficient method of inspecting a complete structure in a short time and has been used, by some offshore operators, for some considerable time now with NDT underwater ultrasonic probes or gamma radiation sources. These schedule-driven inspection techniques require the deployment of a diver or a remotely operated vehicle (ROV)[1]. However, due to the harsh loading environment to which jacket structures are exposed, the time period between crack initiation and the severing of a member is relatively short. This emphasizes the limited life available after through-thickness cracking and raises significant safety issues. This increases financial pressure to lengthen inspection intervals, which could result in a jacket being exposed to a severed member for a considerable duration[2]. The integration of a field-proven underwater NDT technique, such as FMD, within the context of structural health monitoring (SHM), has been proposed by the authors[3]. This may advantageously be applied to overcoming the shortcomings of traditional FMD, by eliminating inspection intervals, providing early warning failure during operation, and bringing about lifecycle cost reduction. The authors have developed a system that employs a single piezoelectric transducer which can be permanently attached to the inner wall of every sub-sea structure and which is powered by a normally inert seawater battery. Upon activation, the sensor transmits ultrasonic chirp or tone encoded pulses, in the range of 21–42 kHz, to a monitoring receiver system at deck level for decoding and identifying flooded members. Guided wave signals have been successfully identified from steel pipes, 7 m in length, 0.5 m in diameter and 16 mm in
thickness, flooded and immersed in seawater. However, the identification of the signals with sufficient Signal to Noise Ratio (SNR) has been performed manually[4]. This work presents an automatic pulse position modulation (PPM) system, which transmits ultrasonic guided waves pulses in the range of 40 kHz, through hollow steel tubes. The system employs an intelligent modulator composed of a PZT element, electronics and batteries. The demodulator instrumentation is made of a PZT ultrasound transducer, a DSP board and a PPM module based on a microcontroller, which performs automatic PPM detection of guided wave energy packets. Experiments carried out in steel tubes, in air, have successfully detected automatically guided wave encoded information. 2. Background theory The frequency equation for wave propagation in a hollow isotropic elastic cylinder has been derived in detail by Gazis[5]. In general, there are three principal types of guided wave mode that can exist in a cylindrical waveguide: the longitudinal (or L(0,m)), the torsional (or T (0,m)) and the flexural (or F(0,m)). In theory, there are an infinite number of individual modes within each principal mode, whose phase velocities, Vph, for a given frequency–thickness product, fd, represent permissible solutions to an implicit transcendental equation of the form Ω M (a, b, λ , µ , f d , V ph ) = 0 (1) where a and b stand for the inner and outer radii of the tube and λ and µ represent its Lame constants. The index M determines the manner in which the fields generated by the guided wave modes vary with the angular coordinate θ in the cylinder cross section. The acoustic fields (i.e. displacement, stress, etc) of the modes when M = 0 are axially symmetric along a cylinder circumference and are independent of the angular coordinate θ. The other modes are nonaxisymmetric and have fields which do vary with the angular coordinate θ. The axisymmetric modes comprise both the longitudinal modes, L(0, n), and the torsional modes, T (0, n); the nonaxisymmetric modes are represented by the flexural modes, F(M, n) [6]. Longitudinal axially symmetric modes are frequently preferred and more implemented in practical applications than torsional and flexural modes. They offer better experimental aspects such as ease of excitability and repeatability than torsional modes, and are preferred over flexural modes for excitation owing to the symmetry that allows the inspection of 360° along the circumference of the tubes. The longitudinal modes are the easiest to generate using conventional wedge type transducers and liquid couplant, compared to the torsional modes, which can only be generated by applying shear forces, their behaviour depends on both the geometry of the tube in which they propagate and the frequency at which they are generated.
1.0
Components
10.0
Vph (m/ms)
8.0
F(1,2)
0.5
0.0
-0.5
6.0 L(0,1)
-1.0
0.0
4.0
1.0
2.0
1.0
2.0
3.0
4.0
3.0
4.0
2.0
Sum
1.0
2.0
F(1,1)
0.0 -1.0 -2.0
0.00
0.02
0.04
0.06
0.08
Frequency-Thickness (MHz-mm)
0.10
0.12
0.0
Time (ms)
b) a) Figure 1. a) Dispersion curve of a hollow steel tube 2mm thickness o.27 m diameter; b) Multiple modes generated from transmission point, including L(0,1), F(1,1) and F(1,2)
Classifications of velocity dispersion curves, which characterize in the solutions to equation (1), are achieved, pragmatically, in terms of a frequency–thickness product, fd, of the waveguide. Figure1a shows the phase velocity dispersion curve generated by the software Disperse[7] of a steel pipe, 2 mm in thickness and 0.27 m in diameter, in air. In these conditions, attenuation of the propagated guided waves packets is considered null. As shown in figure1a, it is possible to have several guided wave modes with a single frequency and all of them with different velocities. The total number of guided wave modes for a given fd value is finite and increases with an increase in frequency. Therefore, having a constant thickness, the selection of a low fd value, where the number of guided wave modes is reduced, depends on a low frequency which possesses a large wavelength. 3. Transducer and frequency selection Guided waves used in long range applications require the employment of frequencies below 100 kHz[8]. The selection of a specific point on the dispersion curves depends not only on the frequency spectrum but also on the phase velocity spectrum associated with the transducer source. Generally, an ultrasonic transducer source can excite all the modes which exist within its frequency spectrum; normally, the spectrum becomes narrower for larger transducers[9]. This is of particular interest considering that PZT elements and an ultrasound transducer procedure with normal beam loading and reception, respectively, have been used in this application. Hence, surface pressure loading will excite longitudinal modes and/or flexural modes depending on the applied pressure distributions on the tube surface[10]. Figure 1b depicts the possible excited modes at 40 kHz according to the software Disperse. As figure 1a indicates, at frequencies beyond 40 kHz (0.08 MHz-mm) the number of the modes increases; hence the selection of the modes becomes difficult. Reducing the frequency of excitation will excite fewer guided wave modes; however, it exposes the overall system to greater risk of audio bandwidth interference, leading to a poor SNR. A compromise between these factors was considered and a transducer with a centre frequency of 40 kHz was selected; thereby, circular PZT elements were powered with tone pulses of 40 kHz. 4. Signal selection Narrow band signals are often used as excitation for NDT purposes in order to give good signal strength and to avoid dispersion over long propagation distances. Tone pulses of between 5 and 10 cycles modulated by a Hanning or a Gaussian window are frequently employed[11]. Nevertheless, experimental results have revealed that narrow band chirp signal and long tone pulses employed in conjunction with rectangular band-pass filtering yield better SNR for this application[12]. Furthermore, it has been shown that the use of square pulses not only increases the energy of the transmitted signals, thereby improving the SNR, but also reduces the complexity of the signal generation hardware[13]. Moreover, it is not necessary to use digital-to-analogue converters to generate square pulses, and power amplifiers can be substituted by high speed switching circuits. In this work, hence, square tone pulses of 40 kHz, 2.5 msec pulse width were implemented. 5. Guided wave PPM system The fundamental research framework of PPM communication systems has been established around 50 years ago[14]; however, these systems have recently experienced important interest with the development of impulse radio and fiber-optic transmission systems[15]. PPM is a form of signal modulation in which the message information is modulated in the time-delay between pulses in a sequence of signal pulses. This method, as opposed to Pulse Width Modulation (PWM) is more power efficient, since PWM employs long pulses, which consumes considerable power, and do not offer additional information[16]. Besides, it has been demonstrated theoretically that PPM systems are effective when the signals are power limited rather than band limited[14].
Steel tube Smart sensor & PPM modulator
Microcontroller based PPM demodulator
FIR digital filter
Figure 2. Guided wave PPM modulator and demodulator communication system
Moreover, the electronics required to demodulate the PPM signals are simple, which leads to small, light-weight receiver/decoder units. Figure 2 depicts a block diagram of a guided wave modulator and demodulator PPM system. In the PPM modulator, constant amplitude pulses are generated at a predefined time delaybetween pulses, where the position of the pulses conveys the signal information. The sensor and modulator design was based on a microcontroller, a current booster, a single PZT element and a 9V battery, as shown in figure 3a. Time-delay
Pulses
Signal Booster
Microcontroller
PZT Element
Channel
Battery Pulse width
a)
Random OFF-TIME
b)
Figure 3. a) PPM modulator main components; b) PPM modulator message information scheme
The design generates 100 square pulses of 40 kHz, i.e. 2.5 ms pulse width. The time-delay between pulses was 45 ms for a sensor 1 and 55 ms for a sensor 2. Each smart sensor was programmed into the internal flash memory of the microcontroller. The digital signal was fed to a current booster and then to a high frequency step-up pulse transformer with an input/output ratio of 1:8. This signal was then applied to the actual PZT element. Figure 3b depicts the modulator message information scheme of the modulator. The PPM demodulator, whose main components are depicted in figure 4, is composed of a 40 kHz ultrasound transducer, a DSP based digital FIR filter, a microcontroller based module, which include an instrumentation amplifier, a threshold comparator, and an USB interface to communicate to a Personal computer. PC
USB
Microcontroller module
Instrumentation Amplifier & Threshold comparator
Digital filter FIR Ultrasound transducer
Figure 4. PPM demodulator principal components
6. Experiment setup An experiment was setup in the laboratory using a steel tube 1.5m in length, 0.27 m in diameter and 2 mm in thickness. A PZT ultrasound transducer, coupled with petroleum jelly, acting as receiver was attached to one extreme of the pipe and two smart sensors and PPM modulators were adhered to opposite extreme and around of the tube using strong magnets. Figure 5 displays this setup.
Figure 5 Experiment setup using a hollow steel pipe, 1.5 m x 0.27 m x 2 mm, two PPM modulators, and the PPM demodulator instrumentation.
7. Results The automatic guided wave PPM system was tested by powering the PPM modulators sensor 1 and sensor 2 simultaneously. The PPM demodulator converts the transmitted signals to a set of pulses by filtering and amplifying, in real-time, guided wave signals, using a sharp FIR filter 2 kHz bandwidth (39 kHz – 41 kHz), an instrumentation amplifier and a threshold comparator embedded in the microcontroller, and this in turn, performs the algorithm to demodulate the original but dispersed PPM pulses. Figure 6 depicts these results. Lower signals are the filtered and amplified guided wave signals and upper signals are the pulses seen by the microcontroller after the threshold comparator. The software verifies that the pulses are 40 kHz, and discriminate the time-delay between pulses allowing 6 ms tolerance due to dispersion; the results are transmitted via USB in ASCII standard to a PC. Collision between modulated transmitted guided waves was minimized by using several off-times in the modulators.
Figure 6. Lower signals are the received guided wave pulses filtered and amplified. Upper signals are the pulses seen by the microcontroller after the threshold comparator.
6. Conclusions A novel automatic PPM guided wave system for the intention of detecting flooded sub-sea members of oil rigs, has been designed, implemented and evaluated. Results show that by exploiting the wave-guide effect of steel pipes acting as communication channel, successful transmissions and reception of encoded PPM information has been attained based on the theoretical results yielded by the software Disperse. Although the tests have been conducted over small distances in air, the feasibility of automatically detecting dispersive guided wave energy packets, provided sufficient SNR, has been proved. These results are very encouraging, taking the authors to the next stage of this work, which is to carry out an experiment in a real offshore platform under construction. References [1] ISO/DIS 19902, 2004 Fixed steel offshore structures [2] A. Nelson et al 2004 Stress redistribution in platform substructures due to primary member damage and its effect on structural reliability Research Report 245 (EQE International Limited for the HSE) [3] R. Mijarez, P. Gaydecki, M. Burdekin, Flood member detection for real-time structural health monitoring of sub-sea structures of offshore steel oilrigs, Smart Materials and structures Vol 16, number 5, October 2007, pp 1857-1869. IOP Publishing, ISSN: 0957-0233. [4] R. Mijarez, P. Gaydecki, M. Burdekin, “An axisymmetric guided wave encoded system for flood detection of oil rig cross-beams” Measurement Science and Technology, Vol. 16 Issue 11 article 019 (2005) IOP Publishing, pp 2265-2274. ISSN: 0957-0233. [5] D. C. Gazis 1959 Three dimensional investigation of the propagation of waves in hollow circular cylinders I. analytical foundation J. Acoust. Soc. Am. 31 568–73 [6] M. G. Silk and K F Bainton 1979 The propagation in metal tubing of ultrasonic wave modes equivalent to Lamb waves Ultrasonics 17 11–9 [7] B. Pavlakovic and M Lowe 2003 Disperse User Manual: a System For Generating Dispersion Curves [8] P. Cawley, M. J. S. Lowe, D N Alleyne, B Pavlakovic and P Wilcox 2003 Practical long range guided wave testing: applications to pipes and rail Mater. Eval. 61 66–74 [9] J.L Rose, Ultrasonic waves in solid media, Cambridge University Press, Cambridge 1999. [10] H. J. Shin and J. L. Rose 1999 Guided waves by axisymmetric and non-axisymmetric surface loading on hollow cylinders Ultrasonics 37 355–63 [11] P. Cawley and D. Alleyne 1996 The use of Lamb waves for the long range inspection of large structures Ultrasonics 34 287–90 [12] R. Mijarez, P. Gaydecki and M. Burdekin 2005 Continuous monitoring guided wave encoded sensor for flood detection of oil rig leg crossbeams Insight 47 748–50 [13] M. Pollakowski and H. Ermert 1994 Chirp signal matching and signal power optimization in pulse-echo mode ultrasonic nondestructive testing IEEE Trans. Ultrason., Ferroelectr. Freq. Control 41 655–9 [14] R. J. McAulay, 1968 Optimal Control Techniques Applied to PPM Signal Design, Information and Control 12, 221-235 [15] P. Azmia et al, 2004 An efficient method for demodulating PPM signals based on Reed–Solomon decoding algorithm, Signal Processing 84 (2004) 1823–1836 [16] S. Haykin , Communication Systems, John Wiley & Sons Inc., 2002.
